Anal. Chem. 1003, 65, 1709-1716
1700
Tryptic Mapping of Recombinant Proteins by Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry Todd M.Billeci and John T. Stults' Protein Chemistry Department, Genentech, Znc., 460 Point Son Bruno Boulevard, South San Francisco, California 94080
Various matrix mixtures have been used for matrix-assisted laser desorption/ionization mass spectrometry to characterize the tryptic maps of recombinant human growth hormone (rhGH) and recombinant human tissue plasminogen activator (rt-PA). Carbohydrate-containing comatrices give improved results over single-component matrices. Of those studied, fucose plus 2,Cdihydroxybenzoic acid (DHB) produced a signal for 24 out of 25, or 96%, of the tryptic peptides of rhGH in a single spectrum. These results were obtained for analyses of as little as 280 fmol of unfractionated material measured in digestion buffer. Analysis of 150 fmol showed a decrease in the relative abundance of higher molecular weight peptides. The incorporation of 5-methoxysalicylic acid (5MSA) as a comatrix in a molar ratio of analyte: fucose:DHB:SMSA = 1:5OO05000:50 gave signals for 45 out of 51 peptides for 4.5 pmol of a tryptic digest of rt-PA, corresponding to 88% of the expected fragments. Unobserved peptides were typically di- and tripeptides. Three glycopeptides were observed with peaks corresponding to the known major glycoforms. The fucose/DHB and 5MSAI DHB comatrices produced significant enhancements in spectral quality over DHB alone, including suppression of matrix peaks, increased ion signal, improved resolution, increased number of useful laser shots per crystal, and minimization of baseline slope. Spectra obtained with fucose/ DHB generally surpassed DHB/5MSA in quality, though both matrix mixtures were clearly superior to neat DHB. Fucose/DHB demonstrated a n increase in tolerance to ionic contaminants by producing a 10-fold reduction in the abundance of (M + Na)+ions. A trimatrix, DHB/5MSA/fucose, produced the highest quality spectra to date, although only marginally better than the fucose/ DHB comatrix. INTRODUCTION Peptide mapping is a valuable tool for the characterization of proteins.1J Detailed analysis of each of the component peptides in a protein permits examination of the primary structure in considerable detail. Mass spectrometry is now one of the most useful techniques for peptide mapping.3~~
* To whom correspondence ehould be addressed. (1) Hancock, W.
5.;Bishop, C. A.; Hearn, M. T. W. Anal. Biochem.
1979,89, 203-212. (2) Garnick, R. L.; Solli,N. J.; Papa, P. A. Anal. Chem. 1988,60,2546 2557. (3) Henzel, W. H.; Bourell, J. H.; Stulta, J. T. Anal. Biochem. 1990, 187,228-233. 0003-2700/93/0365-1709$04.00/0
When combined with conventional reversed-phase highperformance liquid chromatography (HPLC), rapid identification of nearly every peptide in a protein is permitted. For proteins of unknown sequence, mass measurements provide valuable information prior to sequencing by Edman degradation or tandem mass spectrometry.5*6For proteins of known or putative sequence, mass spectrometry permits rapid, detailed confirmation of the sequence and the identification of possible protein modifications. In recent years, matrix-assisted laser desorption/ionization mass spectrometry (MALDI) has emerged as an important technique for the analysis of biomolecules.7-16 Among the samples successfully analyzed are crude biological fluids,12 mixtures of high molecular weight proteinsFJ2 oligosaccharides,l7 nucleic acids,leJQand enzymatic protein digesta.12*23 Characterized by sensitivity in the midferntomole to lowpicomole range, MALDI has a high (>300 kDa) upper mass limit, relative tolerance to many buffer components, and the ability to analyze unfractionated biomolecular mixtures. As a result, the technique shows considerable promise for research, development, and quality control applications in biotechnology. In particular, the technique appears to be well-suited for analysis of peptide mixtures that result from proteolytic protein digestion. A number of other mass spectrometric techniques have been used successfully for the characterization of enzymatic (4) Ling, V.; Guzzetta, A. W.; Canova-Davis,E.; Stulta, J.T.; Hancock, W. S.; Covey, T. R.; Shushan, B. I. Anal. Chem. 1991,63, 2909-2916. (5) Henzel, W. J.; Aswad, D. W.; Stulta, J. T. In Techniques inProtein Chemistry; Hugli, T. E., Ed., Academic Press: San Diego, CA, 1989; pp
127-134. (6) Johnson, R. S.; Walsh, K. A. Protein Sci. 1992,1, 1083-1091. (7) Karas,M.; Bachmann, D.; Bahr, U.; Hillenkamp, F. Znt. J. Mass Spectrom. Zon Processes 1987, 78, 53-86.
(8) Karas,M.; Bahr, U.; Hillenkamp, F. Znt. J. Mass Spectrom. Ion Processes 1989,92, 231-242. (9) Hillenkamp, F.; Karas,M.; Beavie, R.; Chait, B. Anal. Chem. 1991, 63,1193-1203. (10) Hillenkamp,F.; Karas,M. Methods Enzymol. 1990,193,280-294. (11) Karas,M.; Bahr,U. Mass Spectrom. Rev. 1991,10, 335-357. (12) Beavie, R.; Chait, B. Proc. NatZ. Acad. Sci. U.S.A. 1990,87,68736877. (13) Beavie, R.; Chait, B. Rapid Commun.Mass Spectrom. 1989,3, 436-439. (14) Beavis, R.; Chait, B. Anal. Chem. 1990,62, 1836-1840. (15) Karas, M.; Ingendoh, A.; Bahr, U.; Hillenkamp, F. Biomed. Environ. Mass Spectrom. 1989,18, 841-843. (16) Beavis, R. Org. Mass Spectrom. 1992,27, 653-659. (17) Stahl,B.;Steup,M.;Karas,M.;Hillenkamp,F.Anal. Chem. 1991, 63.1463-1466.
(18)Parr, G.; Fitzgerald,M.;Smith,L. Rapid Commun.Mass Spectrom. 1992,6,369-372. (19) Tang, K.; Allman, S.; Chen, C. Rapid Commun.Mass Spectrom. 1992,6, 365-368. (20) Caprioli, R.; Whaley, B. In Techniques in Protein Chemistry ZI; Villafranca, J. J., Ed.; Academic Press: San Diego, CA, 1991;pp 479-510. (21) Schar, M.; Borneen, K. 0.; Gassman,E. Rapid Commun.Mass Spectrom. 1991,5, 319-326. (22) Juhasz,P.;Papayannopouloa,I.A.;Zeng,Ch.;Papov,V.;Biemann,
K.Proceedings of the 40th ASMS Conference on Mass Spectrometry and Allied Topics, Washington, DC, 1992; American Society for Mass Spectrometry: Santa Fe, NM, 1992; pp 1913-1914. (23) Andrews, P. C.; Allen, M. H.; Vestal, M. L.; Nelson, R. W. In Techniques in Protein Chemistry ZZI; Angeletti, R. H., Ed.; Academic Press, Inc.: San Diego, CA, 1992; pp 515-523. 0 1993 American Chemical Society
1710
ANALYTICAL CHEMISTRY, VOL. 65, NO. 13, JULY 1, 1993
digests of proteins. Plasma desorption mass spectrometry (PDMW4 has been successfully applied to the mass measurement of mixtures of proteins and peptides.25-27 Considerable tolerance to high concentrations of solution components (e.g., salts, buffers) has been demonstrated for PDMS by washing samples adsorbed to nitrocellulose,% although frequently with some reduction in the abundance of hydrophilic peptides. Despite its tolerance to many contaminants, PDMS has been most limited by its sensitivity, often requiring tens of picomoles of material.25126 Fast atom bombardment (FABP has also shown considerable utility for the characterization of mixtures of peptides.3033 However, this technique displays significant matrix interference for small molecules (2400 was dramatically reduced, and anomalous
peaks emerged in the background. These effects, which increased considerably for analyses in the 40-50-fmol range, probably originate from the adsorption of larger peptides to the sample vial, the increasing importance of the purity of the analyte at lower concentrations, and the cleanliness of the laser target. A second, more complex mixture, the tryptic peptides of rt-PA, was also examined. A comparison of mass spectra of this digest prepared in neat DHB, DHB/fucose, and DHB/ fucose/SMSA is given in Figure 4. The use of the fucose and fucose/5MSA comatrices produced a considerable reduction in the abundance of matrix ions, improved the sensitivity, and increased the numbers of peptide components observed in a single spectrum. To achieve comparable analyte signal with rt-PA prepared in neat DHB, relatively higher laser fluences were required. However, higher laser fluences introduced undesirable artifacta, such as baseline sloping and dimerization (Figure 4A,B), as observed for preparations of rhGH in neat DHB (Figure 2A). The preparation of samples in the DHB/fucose/5MSA trimatrix (analyte:DHB:fucose:5MSA= 1:5000:5000:50) afforded 15-20 crystals which were spheroid, translucent, and evenly distributed across the target face, although concentrated somewhat at the perimeter. The ionization threshold for rt-PA occurred at a higher laser fluence than that for the rhGH tryptic peptides. The spectrum of tryptic rt-PA obtained with fucose/DHB/5MSA (Figure 4D) is given in Figure 5 with the mass axis expanded for detail. Similar enhancements in performance were noted upon addition of fucose to neat DHB (Figure 4 0 . The ionization threshold for the rt-PA sample prepared in DHB/fucose occurred at the same attenuator setting as that for the DHB/fucose/5MSA trimatrix. In the DHB/fucose/BMSA matrix the resolution at 800 < m/z < lo00 was 160 and at 2000 < m/z < 4000 was 260. A list of tryptic peptides of rt-PA observed by MALDI with DHB/fucose/5MSA is given in Table 111. Of the 51 tryptic peptides expected for rt-PA, 45 produced signals in a single spectrum, which corresponds to 88%of the predicted fragments. Among these, five peptides (T5/T15/T41 and T40/
Table 11. Tryptic rhGH Peptides Observed by MALDI in Two Different Matrices amino acid residue no. matrix DHB/fucosec from to tryptic peptidea calcd massb obsd mass A 1 1 7 9 17 20 39
8 6 8 16 19 38 41
42 65 71 78 95 100 116 128 135 141 146 159 159 168 173 179 179
64 70 77 94 115 115 127 134 140 145 158 167 172 172 178 183 191
T1 T1C' T1C" T2 T3 T4 T5 T6-Tl6 T6 T7
T8 T9 T10 TlOC' T11 T12 T13 T14 T15 T16 T17 T17 + T18 T18 + T19 T20 T20-T21
+ T19
931.1 687.9 262.3 980.2 383.4 2343.6 404.4 3764.2 2617.9 762.8 845.0 2056.5 2263.5 1744.9 1362.5 773.8 693.8 626.7 1490.6 1149.4 147.2 1384.4 1254.4 618.8 1401.6
930.8 688.7 262.7 981.0 384.4 2343.4
-0.3 0.8 0.4 0.8 1.0 -0.2
3762 2616.6 765.0 844.9 2057.5 2263.5 1745.7 1362.7 773.4 694.6 627.8 1490.4 1150.4 145.8 1384.2 1254.2 619.4 1403.3
-2.2 -1.3 2.2 -0.1 1.0 0
0.8 0.2 -0.4 0.8 1.1 -0.2 1.0 -1.4 -0.2 -0.2 0.6 1.7
matrix DHB/5MSAd obsd mass A 930.9
-0.2
980.2 382.5 2343.4
0 -0.9 -0.2
3764.3 2617
0.1 -0.9
844.5 2056.6 2264.2
-0.5 0.1 0.7
1363 , 772.5 693.3
0.5 -1.3 -0.5
1490.2 1149.4 147.3 1384.6 1255 618.7 1401.4
-0.4 0 0.1 0.2 0.6 -0.1 -0.2
obsd in PDMSe
+ + + + ++
F W
+
+ +
+ + 8 ++ + + + + + + + 8 + + + + + + +
-
+ + + + + + + + + +
-
+ + +
*
a A "+" linkage denotes an undigested clip. A --" linkage denotes a disulfide bridge. A 'C" indicates a chymotryptic-like clip. Molecular weights are computed from average isotopic masses. Data taken from Figure 1A. Data taken from Figure 3. e Data taken from Tsarbopoulos et al." f Data taken from Canova-Davis et al.99 8 Detection extremely difficult: low abundance, intermittent.
ANALYTICAL CHEMISTRY, VOL. 65, NO. 13, JULY 1, 1993
600
,400
I
1WO
800
1200
1600
1400
f
’i
2000
2500
3000
3500
4000
4500
5000
5500
m/z
Figure 5. Mass spectrum of 4.5 pmoi of the tryptic digest of reduced
and Scarboxymethylated recombinant human tissue plasminogen activator in a 2,Sdihydroxybenzoicacid/fucose/5-methoxysai~icacid matrix. This spectrum is an expansion of the mass axis of Figure 4D.
T47) could not be resolved due to the limited resolving power of the time-of-flight analyzer. This limitation allowed the unambiguous mass measurement of 78% of the expected fragments. Four peptides (T6, T30, T31, and T46) were only partially resolved from neighboring peaks. Five of the six peptides which did not produce signal were di- and tripeptides. The other peptide, T19, is quite hydrophobic and may have limited solubility under the conditions used. However, the T19 peptide remained absent in the MALDI spectrum when the rt-PA tryptic peptides were dissolved in an aqueous solution of 10% acetonitrile/O.l% TFA. The presence of the T19 peptide was later verified by electrospray ionization. Though the results obtained for the DHB/fucose/SMSA and DHB/fucose were roughly equivalent, the principle advantage of the trimatrix was sensitivity in the m/z
